Course Content

Introduction to TensorFlow

1. Tensors

Welcome to TensorFlow Introduction to Tensors Tensor Properties Applications of Tensors Batches Creating Tensors Data Types Basic Operations: Arithmetic Basic Operations: Linear Algebra Challenge: Creating a Neural Network Layer Transformations Reduction Operations Quiz: TensorFlow Basics

2. Basics of TensorFlow

Gradient Tape Graph Execution Neural Network Implementation Quiz: Conclusion Summary

Creating Tensors

This lesson focuses on the creation of tensors using TensorFlow. TensorFlow provides numerous methods to initialize tensors. By the end of this lesson, you will be proficient in generating tensors for a wide range of applications.

Basic Tensor Initializers

tf.constant(): this is the simplest way to create a tensor. As the name suggests, tensors initialized with this method hold constant values and are immutable;


              12345
            
import tensorflow as tf

# Create a 2x2 constant tensor
tensor_const = tf.constant([[1, 2], [3, 4]])
print(tensor_const)

tf.Variable(): unlike tf.constant(), a tensor defined using tf.Variable() is mutable. This means its value can be changed, making it perfect for things like trainable parameters in models;


              12345
            
import tensorflow as tf

# Create a variable tensor
tensor_var = tf.Variable([[1, 2], [3, 4]])
print(tensor_var)

tf.zeros(): create a tensor filled with zeros;


              12345
            
import tensorflow as tf

# Zero tensor of shape (3, 3)
tensor_zeros = tf.zeros((3, 3))
print(tensor_zeros)

tf.ones(): conversely, this creates a tensor filled with ones;


              12345
            
import tensorflow as tf

# Ones tensor of shape (2, 2)
tensor_ones = tf.ones((2, 2))
print(tensor_ones)

tf.fill(): creates a tensor filled with a specific value;


              12345
            
import tensorflow as tf

# Tensor of shape (2, 2) filled with 6
tensor_fill = tf.fill((2, 2), 6)
print(tensor_fill)

tf.linspace() and tf.range(): these are fantastic for creating sequences;


              123456789
            
import tensorflow as tf

# Generate a sequence of numbers starting from 0, ending at 9
tensor_range = tf.range(10)
print(tensor_range)

# Create 5 equally spaced values between 0 and 10
tensor_linspace = tf.linspace(0, 10, 5)
print(tensor_linspace)

tf.random: generates tensors with random values. Several distributions and functions are available within this module, like tf.random.normal() for values from a normal distribution, and tf.random.uniform() for values from a uniform distribution.

Note

You can also set a fixed seed to obtain consistent results with every random number generation using tf.random.set_seed(). However, be aware that by doing this, you'll receive the same number for any random generation within TensorFlow.

If you wish to achieve consistent numbers for a specific command only, you can provide a seed argument to that command with the desired seed value.


              1234567891011121314
            
import tensorflow as tf

# Set random seed
tf.random.set_seed(72)

# Tensor of shape (2, 2) with random values normally distributed 
# (by default `mean=0` and `std=1`)
tensor_random = tf.random.normal((2, 2), mean=4, stddev=0.5)
print(tensor_random)

# Tensor of shape (2, 2) with random values uniformly distributed 
# (by default `min=0` and `max=1`)
tensor_random = tf.random.uniform((2, 2), minval=-2, maxval=2)
print(tensor_random)

Converting Between Data Structures

TensorFlow tensors can be seamlessly converted to and from familiar Python data structures.

From Numpy Arrays: TensorFlow tensors and Numpy arrays are quite interoperable. Use tf.convert_to_tensor();


              12345678910
            
import numpy as np
import tensorflow as tf

# Create a NumPy array based on a Python list
numpy_array = np.array([[1, 2], [3, 4]])

# Convert a NumPy array to a tensor
tensor_from_np = tf.convert_to_tensor(numpy_array)

print(tensor_from_np)

From Pandas DataFrames: for those who are fans of data analysis with Pandas, converting a DataFrame or a Series to a TensorFlow tensor is straightforward. Use tf.convert_to_tensor() as well;


              12345678910
            
import pandas as pd
import tensorflow as tf

# Create a DataFrame based on dictionary
df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})

# Convert a DataFrame to a tensor
tensor_from_df = tf.convert_to_tensor(df.values)

print(tensor_from_df)

Note

Always ensure that the data types of your original structures (Numpy arrays or Pandas DataFrames) are compatible with TensorFlow tensor data types. If there's a mismatch, consider type casting.

Converting a constant tensor to a Variable: you can initialize a Variable using various tensor creation methods such as tf.ones(), tf.linspace(), tf.random, and so on. Simply pass the function or the pre-existing tensor to tf.Variable().


              12345678910111213
            
import tensorflow as tf

# Create a variable from a tensor
tensor = tf.random.normal((2, 3))
variable_1 = tf.Variable(tensor)

# Create a variable based on other generator
variable_2 = tf.Variable(tf.zeros((2, 2)))

# Display tensors
print(variable_1)
print('-' * 50)
print(variable_2)

To get better at creating tensors, try practicing with different shapes and values. For more details on specific commands, check the official TensorFlow documentation. It has all the information you need about any command or module in the library.

Task

Swipe to start coding

Alright! Let's put your newfound knowledge of tensor creation and conversion to the test. Here's your challenge:

Tensor Initialization
- Create a tensor named tensor_A with shape (3,3) having all elements as 5;
- Create a mutable tensor named tensor_B with shape (2,3) initialized with any values of your choice;
- Create a tensor named tensor_C with shape (3,3) filled with zeros;
- Create a tensor named tensor_D with shape (4,4) filled with ones;
- Create a tensor named tensor_E which has 5 linearly spaced values between 3 and 15;
- Create a tensor named tensor_F with random values and shape (2,2).
Conversions
- Given the NumPy array np_array, convert it into a TensorFlow tensor named tensor_from_array;
- Convert the DataFrame df into a tensor named tensor_from_dataframe.

Note

Make sure to use the most appropriate commands for the situation (e.g., create an array filled with ones using tf.ones() rather than tf.fill()).

Solution

Switch to desktop for real-world practiceContinue from where you are using one of the options below

Everything was clear?

Thanks for your feedback!

Section 1. Chapter 6

Creating Tensors

Basic Tensor Initializers

tf.constant(): this is the simplest way to create a tensor. As the name suggests, tensors initialized with this method hold constant values and are immutable;


              12345
            
import tensorflow as tf

# Create a 2x2 constant tensor
tensor_const = tf.constant([[1, 2], [3, 4]])
print(tensor_const)

tf.Variable(): unlike tf.constant(), a tensor defined using tf.Variable() is mutable. This means its value can be changed, making it perfect for things like trainable parameters in models;


              12345
            
import tensorflow as tf

# Create a variable tensor
tensor_var = tf.Variable([[1, 2], [3, 4]])
print(tensor_var)

tf.zeros(): create a tensor filled with zeros;


              12345
            
import tensorflow as tf

# Zero tensor of shape (3, 3)
tensor_zeros = tf.zeros((3, 3))
print(tensor_zeros)

tf.ones(): conversely, this creates a tensor filled with ones;


              12345
            
import tensorflow as tf

# Ones tensor of shape (2, 2)
tensor_ones = tf.ones((2, 2))
print(tensor_ones)

tf.fill(): creates a tensor filled with a specific value;


              12345
            
import tensorflow as tf

# Tensor of shape (2, 2) filled with 6
tensor_fill = tf.fill((2, 2), 6)
print(tensor_fill)

tf.linspace() and tf.range(): these are fantastic for creating sequences;


              123456789
            
import tensorflow as tf

# Generate a sequence of numbers starting from 0, ending at 9
tensor_range = tf.range(10)
print(tensor_range)

# Create 5 equally spaced values between 0 and 10
tensor_linspace = tf.linspace(0, 10, 5)
print(tensor_linspace)

tf.random: generates tensors with random values. Several distributions and functions are available within this module, like tf.random.normal() for values from a normal distribution, and tf.random.uniform() for values from a uniform distribution.

Note

You can also set a fixed seed to obtain consistent results with every random number generation using tf.random.set_seed(). However, be aware that by doing this, you'll receive the same number for any random generation within TensorFlow.

If you wish to achieve consistent numbers for a specific command only, you can provide a seed argument to that command with the desired seed value.


              1234567891011121314
            
import tensorflow as tf

# Set random seed
tf.random.set_seed(72)

# Tensor of shape (2, 2) with random values normally distributed 
# (by default `mean=0` and `std=1`)
tensor_random = tf.random.normal((2, 2), mean=4, stddev=0.5)
print(tensor_random)

# Tensor of shape (2, 2) with random values uniformly distributed 
# (by default `min=0` and `max=1`)
tensor_random = tf.random.uniform((2, 2), minval=-2, maxval=2)
print(tensor_random)

Converting Between Data Structures

TensorFlow tensors can be seamlessly converted to and from familiar Python data structures.

From Numpy Arrays: TensorFlow tensors and Numpy arrays are quite interoperable. Use tf.convert_to_tensor();


              12345678910
            
import numpy as np
import tensorflow as tf

# Create a NumPy array based on a Python list
numpy_array = np.array([[1, 2], [3, 4]])

# Convert a NumPy array to a tensor
tensor_from_np = tf.convert_to_tensor(numpy_array)

print(tensor_from_np)

From Pandas DataFrames: for those who are fans of data analysis with Pandas, converting a DataFrame or a Series to a TensorFlow tensor is straightforward. Use tf.convert_to_tensor() as well;


              12345678910
            
import pandas as pd
import tensorflow as tf

# Create a DataFrame based on dictionary
df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})

# Convert a DataFrame to a tensor
tensor_from_df = tf.convert_to_tensor(df.values)

print(tensor_from_df)

Note

Always ensure that the data types of your original structures (Numpy arrays or Pandas DataFrames) are compatible with TensorFlow tensor data types. If there's a mismatch, consider type casting.

Converting a constant tensor to a Variable: you can initialize a Variable using various tensor creation methods such as tf.ones(), tf.linspace(), tf.random, and so on. Simply pass the function or the pre-existing tensor to tf.Variable().


              12345678910111213
            
import tensorflow as tf

# Create a variable from a tensor
tensor = tf.random.normal((2, 3))
variable_1 = tf.Variable(tensor)

# Create a variable based on other generator
variable_2 = tf.Variable(tf.zeros((2, 2)))

# Display tensors
print(variable_1)
print('-' * 50)
print(variable_2)

Task

Swipe to start coding

Alright! Let's put your newfound knowledge of tensor creation and conversion to the test. Here's your challenge:

Tensor Initialization
- Create a tensor named tensor_A with shape (3,3) having all elements as 5;
- Create a mutable tensor named tensor_B with shape (2,3) initialized with any values of your choice;
- Create a tensor named tensor_C with shape (3,3) filled with zeros;
- Create a tensor named tensor_D with shape (4,4) filled with ones;
- Create a tensor named tensor_E which has 5 linearly spaced values between 3 and 15;
- Create a tensor named tensor_F with random values and shape (2,2).
Conversions
- Given the NumPy array np_array, convert it into a TensorFlow tensor named tensor_from_array;
- Convert the DataFrame df into a tensor named tensor_from_dataframe.

Note

Make sure to use the most appropriate commands for the situation (e.g., create an array filled with ones using tf.ones() rather than tf.fill()).

Solution

Switch to desktop for real-world practiceContinue from where you are using one of the options below

Everything was clear?

Thanks for your feedback!

Section 1. Chapter 6

Switch to desktop for real-world practiceContinue from where you are using one of the options below